Date: Tue, 3 Oct 2000 15:10:02 -0400 (EDT) From: Garrett Wollman <wollman@khavrinen.lcs.mit.edu> To: Bruce Evans <bde@zeta.org.au> Cc: freebsd-bugs@FreeBSD.ORG, freebsd-standards@khavrinen.lcs.mit.edu Subject: Re: misc/21644: /usr/include/sys/mman.h uses a type defined in /usr/include/sys/types.h Message-ID: <200010031910.PAA43095@khavrinen.lcs.mit.edu> In-Reply-To: <Pine.BSF.4.21.0009301016180.3321-100000@besplex.bde.org> References: <200009291628.MAA92746@khavrinen.lcs.mit.edu> <Pine.BSF.4.21.0009301016180.3321-100000@besplex.bde.org>
next in thread | previous in thread | raw e-mail | index | archive | help
<<On Sat, 30 Sep 2000 10:31:08 +1100 (EST), Bruce Evans <bde@zeta.org.au> said: > Not precisely the change in the PR. <sys/types.h> has too much namespace > pollution to include nested. The correct change (unless P1003.1-200x is > so broken as to require everything that may be in a POSIX <sys/types.h> > to be declared in <sys/mman.h>) is to replace mode_t by a basic type that > is compatible with mode_t and whose name is not in the application namespace. In certain instances, POSIX actually requires that the types themselves be defined; for example, <sys/stat.h> is required to cause the definition of the types `blkcnt_t', `blksize_t', `dev_t', `ino_t', `mode_t', `nlink_t', `uid_t', `gid_t', `off_t', and `time_t' from <sys/types.h>. This suggests to me that we may need to use the _BSD_SIZE_T trick on many or all of the types in question. If we can centralize these definitions it would be better than allowing the various _t types to potentially become desynchronized with their consumers. -GAWollman To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-bugs" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200010031910.PAA43095>